So What Is The Pandora Battery Life?

[Dimacus]

I just want to point out that the wireless mice that boast with only drawing X mA only refer to the mouse itself, the adapter however might draw the full 5v 500mA, which means 2.5W.
Now, doing some calculations based on some guesswork and the numbers mweston specified, we have a ~4000mAh battery driving the pandora for ~8.5 hour.
This tells us that the current being drawn by the whole system is about ~470mA at 3.6V (I'm guessing the battery is providing this voltage normally).
(At this point I'm starting to doubt my numbers, since the screen is taking about 1.6W, that leavs 100mW for the rest of the system, which, if I am reading the omap data sheet correctly, is to little ).
So the system draws about 1.7W(!) in total. (Compare that to the 2.5Watts of power for the full USB port.)
So, adding something powerhungry, like some cheap USB mice or keyboards can eat alot of power.


*Edit*
Bah, that's what you get for being social with your friends and forgetting to press the post button.

Re-reading MWeston's posts, I think I read It a little to fast, and therefor incorrectly assumed the LCD consumes about 1.6W, and not the full system with the LCD at max brightness.
I was wondering about that, now it makes a whole lot more sense.

 

[God Ginrai]

EDIT: Hey, why aren't my posts being combined together?

I don't think that has ever been a feature on GP32X. I have seen a lot of people double post with some sort of ill idea that the forum does this, but the only times I've ever seen posts merged together on any forum is when a mod takes the time to do it. I mean hell, the forums give you a flood warning if you try to post again too soon after your last posts.

-God Ginrai

 

[gp2.eXe]

Soooo... we're talking that the Pandora's battery life could be decreased from around 10 hours to less than 5? I get why you're concerned...

To me? Nope, it is like meh, it's the lesser of my worries. I'm happy with my '2X lasting only 2 hours!

(Yes, I need to get a new pair of batteries, but I'm saving for Pandora, mwahaha.)

 

[Pleng]

So what's going to be a bigger drain on power: Keeping bluetooth on all the time and using a bluetooth mouse, or using a USB mouse?

 

[Kramy]

The CPU uses about 0.05w, and the LCD about 1.6w. Downclocking will knock that down to 0.04w, which doesn't save a whole heck of a lot of battery life. ;"]

Where on earth did you get such terribly inaccurate numbers!?[/quote]The LCD uses 0.5W for the backlight and maybe 0.1W for the LCD logic.
It was this guy called MWeston, that posted in this thread.
QUOTE
Average power with the LCD at full brightness is around 1.6W. Adding the SGX core into the mix (like with Quake) added another 30-60mA of drain I think.

Perhaps I misunderstood? Sorry.

 

[VRAndy]

I think you did. "Average power with the LCD", not "Average power of the LCD".

 

[Monk]

Wireless mice are designed to be power misers, since the designers want to maximise battery lifetime. But that doesn't mean normal mice/keyboards are - it's easy for the designer to be lazy or cheap and suck up whatever power is convenient, instead of spending extra time optimising the design or paying for more expensive low-power components.

Makes sense, though the apparent degree was surprising and the concept shows just how wasteful we TEND to be of resources

 

[Monk]

EDIT: Hey, why aren't my posts being combined together?

I don't think that has ever been a feature on GP32X. I have seen a lot of people double post with some sort of ill idea that the forum does this, but the only times I've ever seen posts merged together on any forum is when a mod takes the time to do it. I mean hell, the forums give you a flood warning if you try to post again too soon after your last posts.

-God Ginrai

Yes, it did used to happen. Quite surprised me the first time as it doesn't happen on other forums I frequent.

 

[dingens]

So what's going to be a bigger drain on power: Keeping bluetooth on all the time and using a bluetooth mouse, or using a USB mouse?

Bluetooth should be much better. Pandora has got a Class 2 transceiver with a maximum transmit power of 2.5mW. There will be loss while generating 2.5mW transmit power but it should be far less than wired devices need(500mW for a mouse for example, 100mA @ 5V). Bluetooth transceivers also have got powersaving features that reduce the consumption further when idle.

 

[aq21]

The OMAP takes more than 500mW when active, I don't know where the 50mW figure comes from. See page 6 of this document:

http://focus.ti.com/lit/an/sprt495/sprt495.pdf

The 540mW figure excludes some of the chip functions (I/O, which is not trivial, being at a higher voltage, for example), and is with SGX not working. All things counted, and power conversions inefficiencies accounted for, the OMAP itself will take about 1W from the battery while playing Quake. If the screen takes about 1.5W, then there is no way for 10 hours of play. Even 5 hours sounds very optimistic.

 

[dingens]

If the screen takes about 1.5W, then there is no way for 10 hours of play. Even 5 hours sounds very optimistic.

But it doesn't. See above

 

[Polossatik]

I think it would be a good idea if someone (preferably MWeston ) could fork out the time to do some basic testing and make a nice overview of what draws how much power.
Something Like - (maybe better idea's can be provided)

wifi/BT/LCD off - idle system totally in "nothing happens modus" just Linux kernel booted and making once every time a few cpu cycles
wifi/BT/LCD off - system in "half cpu power mode (read = MP3 player)"
wifi/BT/LCD off - system in "full cpu mode (read = non-optimised apps (dosbox?) running)"
wifi/BT/LCD off - system in "full cpu mode with DSP on (read = pandora-optimised apps (quake?) running)"

and then the exact power figures how much LCD on (full/mid/dimmed bright/ lcd on but no backlight), BT on, Wifi on, access SD card/ <insert anything I've forgotten here>, etc uses

This will give a very good idea about power consumption, and then it will also make things very clear about where to do the first efforts for getting decent power saving.

for example I assume that it does make a LOT more sense to have a easy way to configure "LCD off/dimm everything down " (maybe even make it application based possible - so that with MP3 in front/full screen this app can kill the LCD after 20 sec , while otherwise it would wait 3 min or so (=browsing/"system wide" setting)) then trying to make a MP3 player using the DSP alone

Edit: for example on Nokia N800 it's quite annoying that the standard menu has only a separate option to "lock screen" and "switch off BT/WIFI" and no combined ""lock screen AND switch off BT/WIFI". Taking proper care of little things (or allowing people to configure it to have this) like this make much difference then trying to do funcky stuff. (I do realise this is not exactly applicable to to Pandora as such, but it's an example)

 

[cosurgi]

, you get the 10 hour life and if I turned down the LCD brightness it would be a bit more.

Thanks MWeston for nice response. So LCD & USB are affecting a ~9h battery life by roughly 25% (and up to 50% in case of power hungry USB things).

How about wifi and BT? Does keeping them turned on/off has impact smaller than 10% ?

best regards

 

[dingens]

"

I don't think that it makes much sense to create such a table without including proper powers saving software in the kernel first. A software-side implementation is necessary to use the SoC to its full potential(switching off/ideling components/cores, lowering core-clock/Vcore). There is surely a reason why Nokia doesn't open source their power-saving module.
Quake doesn't use the DSP afaik.
Loading dummy-software, that maxes out every core would give a nice worst-case figure(while still being far away from a normal usage scenario, though).

 

[Eversmile]

EDIT: Hey, why aren't my posts being combined together?

I don't think that has ever been a feature on GP32X. I have seen a lot of people double post with some sort of ill idea that the forum does this, but the only times I've ever seen posts merged together on any forum is when a mod takes the time to do it. I mean hell, the forums give you a flood warning if you try to post again too soon after your last posts.

-God Ginrai

Sometimes, I like to bold things too. Still, I think the forum should do that. I mean, everyone complains when someone double posts, so why do they allow it? I've never coded/managed a forum before, but it can't be that hard, especially since every account has a blog/loads of special settings/other complex stuff. (and this forum is so lightning quick!)

How about wifi? How much does that eat? I mean, if I turn on wifi on my Mr. iPod Touch, the poor thing becomes powerless within 2-4 hours.

 

[(naw)mcx]

EDIT: Hey, why aren't my posts being combined together?

I don't think that has ever been a feature on GP32X. I have seen a lot of people double post with some sort of ill idea that the forum does this, but the only times I've ever seen posts merged together on any forum is when a mod takes the time to do it. I mean hell, the forums give you a flood warning if you try to post again too soon after your last posts.

-God Ginrai

Yes, it did used to happen. Quite surprised me the first time as it doesn't happen on other forums I frequent.

Yeah, it did, I remember that was the best way to add pictures to your post, as editing fucked up everything.

 

[Pleng]

So what's going to be a bigger drain on power: Keeping bluetooth on all the time and using a bluetooth mouse, or using a USB mouse?

Bluetooth should be much better. Pandora has got a Class 2 transceiver with a maximum transmit power of 2.5mW. There will be loss while generating 2.5mW transmit power but it should be far less than wired devices need(500mW for a mouse for example, 100mA @ 5V). Bluetooth transceivers also have got powersaving features that reduce the consumption further when idle.


OK That's encouraging.

Pardon my ignorance on the Bluetooth issue, but is there a heavier load on bluetooth when more devices are being used simultaneously?

I know that having bluetooth turned on means the device draws more power than turned off. But if, for example I am already using bluetooth to connect a mouse, would I suffer *additional* overhead if I also used a bluetooth keyboard? And then tethered a mobile phone for an internet connection?

 

[dingens]

OK That's encouraging.

Pardon my ignorance on the Bluetooth issue, but is there a heavier load on bluetooth when more devices are being used simultaneously?

I know that having bluetooth turned on means the device draws more power than turned off. But if, for example I am already using bluetooth to connect a mouse, would I suffer *additional* overhead if I also used a bluetooth keyboard? And then tethered a mobile phone for an internet connection?

2.5mW is the maximum transmit power regardless how many devices are connected. The difference would be, that the bluetooth module would have less idle moments, I suppose. I don't know, how connecting several devices works with bluetooth. I think the transmitter would permanently cycle through the connections, because it can't establish multiple connections at once. That could mean lower throughput when using tethering for example if other devices have to be fed with data also, like a headset or so. I'm no bluetooth expert, so this is just speculation.

 

[Gren]

On a battery related topic........(I had a quick search of the forums but could't find an answer to this).

On the pandora what is/will be the recommended discharge/charge cycle?
Charge to 100% and drain to 0% before charging again?
If it isn't used for a month or 2 should it be drained or charged before "storage"?
Will there be any long term damage to the battery if it's always plugged into the mains (my work notebook now lasts for 5 mins 'cos it's always docked at work).

I obviously wanna keep the battery in tip-top condition for when it's out and about for extended periods

Apologies in advance to the sensitive types if this has been asked/answered before. I haven't been here for a few months, and Jacquelyn's email has prompted me to get excited again

 

[fusion_power]

, adding 500mA from the USB port to the mix means a 50% reduction in battery life. You don't get that much power for free! That's an extra 2.5W and even at 95% efficiency, it's a lot on the battery.

Wow, scary. ""
I really thought we could Play i.E. Quake 3 with full SGX acceleration for around 8-10 hours. Now it sounds more like we will have a lifetime like the GP2X. Hmmm.... actualy, I mainly ordered the Pandora for the huge Battery-life so I really hope, there will be much optimizing work as soon as possible. If TI says, the OMAP is such a power-saving-wonder-chip I hope they really mean it. ^_^

Reality can be very hard sometimes... my little Microsoft Intelli Mouse (optical) is not designed to be onto a Pandora, onto the Mouse it is printed "100mA" hmm....

How much Power needs the SGX Unit of the Pandora under full load?

The OMAP takes more than 500mW when active, I don't know where the 50mW figure comes from. See page 6 of this document:

http://focus.ti.com/lit/an/sprt495/sprt495.pdf

The 540mW figure excludes some of the chip functions (I/O, which is not trivial, being at a higher voltage, for example), and is with SGX not working. All things counted, and power conversions inefficiencies accounted for, the OMAP itself will take about 1W from the battery while playing Quake. If the screen takes about 1.5W, then there is no way for 10 hours of play. Even 5 hours sounds very optimistic.

The document is quite interesting. At least it sounds good, especialy the part with the low power usage during audio decoding (22mW)
But do anything of this stuff actualy work onto the Pandora? Or is this (automatic???) hardware power managemant currently "overwritten" with unoptimized Software?